Automated pressure ulcer dimension measurements using a depth camera.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Artificial Intelligence in Healthcare

Background:

• Accurate pressure ulcer (PU) monitoring is crucial for effective wound care.
• Current methods often involve manual measurements, which can be time-consuming and subjective.
• Automated, non-contact measurement systems can enhance monitoring accuracy and efficiency.

Purpose of the Study:

• To develop an automatic wound segmentation method for a pressure ulcer monitoring system (PrUMS).
• To provide automated, non-contact wound measurements using a depth camera.
• To improve PrUMS accuracy by combining multiple convolutional neural network classifiers.

Main Methods:

• Developed an automatic wound segmentation method using multiple convolutional neural network classifiers.
• Integrated the method into a pressure ulcer monitoring system (PrUMS) with a depth camera.
• Compared PrUMS measurements with standardized manual measurements from wound care nurses.

Main Results:

• The automatic segmentation method showed measurement errors of 9.27mm (length), 5.89mm (width), and 5.79mm (depth) compared to ground truth.
• Semi-automatic segmentation yielded smaller errors: 4.72mm (length), 4.34mm (width), and 5.71mm (depth).
• No significant differences were found for length and width measurements between segmentation methods and ground truth, but depth measurements differed significantly (p<0.001).

Conclusions:

• The novel PrUMS device offers objective, non-contact wound measurement suitable for clinical practice.
• Using regular camera images can enhance classifier performance.
• While length and width measurements were statistically similar to manual ones (p>0.05), depth measurements showed a statistical difference (p=0.04) due to depth camera limitations.